Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Predicting Time Pressure of Powered Two-Wheeler Riders for Proactive Safety Interventions

Sumit S. Shevtekar, Chandresh K. Maurya, Gourab Sil, Subasish Das

TL;DR

This work addresses the problem that time pressure (TP) degrades powered two-wheeler safety and that TP is difficult to measure in real time. It introduces a large-scale labeled simulator dataset (129,000+ samples, 153 rides, 51 participants) and the MTPS architecture, which combines Conv1D feature extraction, dual-stage temporal attention, and Squeeze-and-Excitation blocks to predict TP states (HTP, LTP, NTP) with high accuracy (91.53%) and ROC-AUC (98.93%). The authors demonstrate that TP-prediction features improve downstream collision-risk prediction from 91.25% to 93.51%, approaching an oracle (93.72%), and show MTPS is compact enough for edge deployment (172k parameters, 0.34 M FLOPs, 2.16 MB). By mapping TP states to graded ITS interventions under the Safe System Approach, the work establishes a practical pathway for proactive safety management in PTW mobility. Overall, MTPS provides a data-driven, real-time cognitive-state predictor that enhances proactive safety interventions and opens avenues for real-world validation and multimodal extensions.

Abstract

Time pressure critically influences risky maneuvers and crash proneness among powered two-wheeler riders, yet its prediction remains underexplored in intelligent transportation systems. We present a large-scale dataset of 129,000+ labeled multivariate time-series sequences from 153 rides by 51 participants under No, Low, and High Time Pressure conditions. Each sequence captures 63 features spanning vehicle kinematics, control inputs, behavioral violations, and environmental context. Our empirical analysis shows High Time Pressure induces 48% higher speeds, 36.4% greater speed variability, 58% more risky turns at intersections, 36% more sudden braking, and 50% higher rear brake forces versus No Time Pressure. To benchmark this dataset, we propose MotoTimePressure, a deep learning model combining convolutional preprocessing, dual-stage temporal attention, and Squeeze-and-Excitation feature recalibration, achieving 91.53% accuracy and 98.93% ROC AUC, outperforming eight baselines. Since time pressure cannot be directly measured in real time, we demonstrate its utility in collision prediction and threshold determination. Using MTPS-predicted time pressure as features, improves Informer-based collision risk accuracy from 91.25% to 93.51%, approaching oracle performance (93.72%). Thresholded time pressure states capture rider cognitive stress and enable proactive ITS interventions, including adaptive alerts, haptic feedback, V2I signaling, and speed guidance, supporting safer two-wheeler mobility under the Safe System Approach.

Predicting Time Pressure of Powered Two-Wheeler Riders for Proactive Safety Interventions

TL;DR

This work addresses the problem that time pressure (TP) degrades powered two-wheeler safety and that TP is difficult to measure in real time. It introduces a large-scale labeled simulator dataset (129,000+ samples, 153 rides, 51 participants) and the MTPS architecture, which combines Conv1D feature extraction, dual-stage temporal attention, and Squeeze-and-Excitation blocks to predict TP states (HTP, LTP, NTP) with high accuracy (91.53%) and ROC-AUC (98.93%). The authors demonstrate that TP-prediction features improve downstream collision-risk prediction from 91.25% to 93.51%, approaching an oracle (93.72%), and show MTPS is compact enough for edge deployment (172k parameters, 0.34 M FLOPs, 2.16 MB). By mapping TP states to graded ITS interventions under the Safe System Approach, the work establishes a practical pathway for proactive safety management in PTW mobility. Overall, MTPS provides a data-driven, real-time cognitive-state predictor that enhances proactive safety interventions and opens avenues for real-world validation and multimodal extensions.

Abstract

Time pressure critically influences risky maneuvers and crash proneness among powered two-wheeler riders, yet its prediction remains underexplored in intelligent transportation systems. We present a large-scale dataset of 129,000+ labeled multivariate time-series sequences from 153 rides by 51 participants under No, Low, and High Time Pressure conditions. Each sequence captures 63 features spanning vehicle kinematics, control inputs, behavioral violations, and environmental context. Our empirical analysis shows High Time Pressure induces 48% higher speeds, 36.4% greater speed variability, 58% more risky turns at intersections, 36% more sudden braking, and 50% higher rear brake forces versus No Time Pressure. To benchmark this dataset, we propose MotoTimePressure, a deep learning model combining convolutional preprocessing, dual-stage temporal attention, and Squeeze-and-Excitation feature recalibration, achieving 91.53% accuracy and 98.93% ROC AUC, outperforming eight baselines. Since time pressure cannot be directly measured in real time, we demonstrate its utility in collision prediction and threshold determination. Using MTPS-predicted time pressure as features, improves Informer-based collision risk accuracy from 91.25% to 93.51%, approaching oracle performance (93.72%). Thresholded time pressure states capture rider cognitive stress and enable proactive ITS interventions, including adaptive alerts, haptic feedback, V2I signaling, and speed guidance, supporting safer two-wheeler mobility under the Safe System Approach.
Paper Structure (52 sections, 14 equations, 8 figures, 15 tables)

This paper contains 52 sections, 14 equations, 8 figures, 15 tables.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Time Pressure and Driving Behaviour
  4. Time Pressure and Gap-Acceptance Behaviour
  5. Time Pressure Effects on PTW Riding: Speed and Over-Speeding
  6. Time Pressure and Distraction in MTW Riders
  7. Simulator Validity Under Time Pressure
  8. Machine Learning for Behaviour and Risk Modeling
  9. Research gap and motivation
  10. Methodology for Data Collection
  11. Simulator Setup
  12. Simulator Scenario Design
  13. Participants
  14. TP Scenarios
  15. Behavioral Transition Under Exam Pressure: Why LTP Matters
  16. ...and 37 more sections

Figures (8)

  • Figure 1: Registered motor vehicles in India by category (2003–2022) MORTH202425.
  • Figure 2: Sample snapshots of the simulated riding environment.
  • Figure 3: Sample participant snapshots from two-wheeler simulator sessions.
  • Figure 4: Flow of dataset collection and experimental design.
  • Figure 5: MotoTimePressure (MTPS) architecture for predicting rider time-pressure states.
  • ...and 3 more figures